一种获得非线性生长函数参数初始值的新方法:杂种遗传算法  

A New Method for Getting Starting Values of Parameters in Nonlinear Growth Functions:Hybrid Genetic Algorithm

作　　者：张元跃[1] 

机构地区：[1]湖南农业大学动物科学技术学院,湖南长沙410128

出　　处：《生物数学学报》2010年第4期731-739,共9页Journal of Biomathematics

基　　金：湖南省科技厅项目(2007NK3135)

摘　　要：本研究旨在杂种遗传算法应用于非线性生长函数的参数估计.提出了杂种遗传算法估计非线性生长函数参数的数学模型.5种非线性生长函数Gompertz、Logistic、von Bertalanffy、Richards、Brody分别拟合一个较大型的、群体类型差异大的番鸭体重生长资料,利用杂种遗传算法获得了有效初始值,在lsqcurvefit与proc nlin中获得了一致最优解的结果.表明杂种遗传算法估计非线性函数参数的实际可行性.This study was conducted to apply hybrid genetic algorithm to the estimation of the starting values of nonlinear growth functions.The mathematical model of hybrid genetic algorithm to estimate parameters in nonlinear growth functions were presented.The 5 nonlinear growth functions,Gompertz,Logistic,von Bertalanffy,Richards,and Brody,were proposed respectively to fit the larger-scale data from body weight of native Muscovy with the large difference of flock type,obtain their effective starting values using hybrid genetic algorithm,and consistently converge at the optimal point using lsqcurvefit and proc nlin procedures.These showed that hybrid genetic algorithm is feasible for optimization problems of nonlinear growth functions.

关 键 词：非线性生长函数 参数估计 杂种遗传算法 番鸭 体重 

分 类 号：Q332[生物学—遗传学]